Qian Dong comments

Results 17 comments of


                                            Qian Dong

How to run the silver annotation pipeline

(core dumped)

Shared tensors not correctly saved.

i have the same issue

Shared tensors not correctly saved.

I run the same code with the following codes and got the issue. https://github.com/huggingface/trl/issues/1121

[BUG] batch_size check failed with zero 2 (deepspeed v0.9.0)

same problem with zero 2 and deepspeed=0.9.0

Turn it to windows

Hi ,i turn it to VS2015,you can see it on https://github.com/DQ0408/cpu_tsdf if you like it, give me a star THX..😄 I will use Linux next time. Windows is too.........

Turn it to windows

@sdmiller No problem, I will turn it into a Pull Request after perfecting this progect😄

[Question] How to improve efficiency for TileLang kernel computing torch.einsum('thd,kd->thk', query, key).sum(dim=1)

```python import torch import tilelang import tilelang.language as T @tilelang.jit( out_idx=[-1], ) def kernel_impl( t: int, h: int, d: int, threads: int, dtype: str = "bfloat16", accum_dtype: str = "float",...